microsoft / microsoft/Mage

I implemented ComfyUI support for Mage Flow, but the results are underwhelming

Open
#17 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.6k
Forks
174
PR merge metrics
No merged PRs in 30d

Description

The edit model lacks reference-based generation capabilities, and the character output from the three T2I models is underwhelming. The image editing functionality is passable, but it doesn't seem to offer any significant advantages.

Specifically, the Turbo model performs poorly because it doesn't allow negative prompts at CFG 1. Its multi-reference capability is virtually non-existent. While the non-Turbo models produce fewer errors, they still hold no competitive advantage over other open-source models. Furthermore, the image editing feature tends to erase details from the original image.

https://github.com/wuwukaka/ComfyUI-Mage
https://huggingface.co/wuwukasi/mage-flow-comfyui

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing the ComfyUI implementation and model links in the issue, then reproduce the reported Turbo, multi-reference, and image-editing behavior. The issue does not identify repository files, tests, or a measurable acceptance criterion, so the expected output and definition of done need to be clarified before implementation.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface, python
Domain
machine-learning
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.